Symmetric Inclusion-Exclusion

One form of the inclusion-exclusion principle asserts that if A and B are functions of finite sets then the formulas A(S) = ∑ T⊆S B(T ) and B(S) = ∑ T⊆S(−1) |S|−|T A(T ) are equivalent. If we replace B(S) by (−1)B(S) then these formulas take on the symmetric form A(S) = ∑ T⊆S (−1) B(T ) B(S) = ∑ T⊆S (−1) A(T ). which we call symmetric inclusion-exclusion. We study instances of symmetric inclusion-exclusion in which the functions A and B have combinatorial or probabilistic interpretations. In particular, we study cases related to the Pólya-Eggenberger urn model in which A(S) and B(S) depend only on the cardinality of S. 1. Inclusion-exclusion Let A and B be two functions defined on a set D of finite sets. We assume that if S ∈ D and T ⊆ S then T ∈ D. Then the inclusion-exclusion principle asserts that the following are equivalent: A(S) = ∑